Related Experiment Video
Updated: Jul 4, 2026
06:04
Systematic Hearing Performance Evaluation Process for Adolescents with Cochlear Implantation at Early Ages
Published on: March 24, 2023
Effect of spectral normalization on different talker speech recognition by cochlear implant users
Chuping Liu1, John Galvin, Qian-Jie Fu
1Department of Electrical Engineering, University of Southern California, Los Angeles, California 90089, USA. chupingl@usc.edu
The Journal of the Acoustical Society of America
|June 6, 2008
Summary
Spectral normalization improves speech understanding for cochlear implant (CI) users by adjusting for talker variations. This technique enhances intelligibility, particularly with challenging acoustic conditions and diverse speakers.
Area of Science:
- Audiology
- Speech Processing
- Biomedical Engineering
Background:
- Cochlear implants (CIs) present challenges in speech understanding due to spectral distortion.
- Variability in talker characteristics (e.g., fundamental frequency, vocal tract) impacts CI user performance.
Purpose of the Study:
- To investigate the effectiveness of spectral normalization in improving speech recognition for cochlear implant users.
- To assess the impact of spectral normalization on speech intelligibility across different talkers and simulated vocal variations.
Main Methods:
- Experiment 1: Measured speech recognition in CI users with and without spectral normalization using two talkers.
- Experiment 2: Assessed spectral normalization effects in CI users and normal-hearing (NH) subjects using pitch-shifting to simulate diverse talkers.
- Applied a spectral normalization algorithm to transform spectral characteristics between talkers.
Main Results:
- Spectral normalization showed a small but significant improvement in speech recognition for CI users in Experiment 1.
- In Experiment 2, spectral normalization significantly enhanced the intelligibility of simulated talkers for CI subjects, despite intersubject variability.
- NH subjects demonstrated near-perfect performance with pitch-shift transformations, serving as a baseline.
Conclusions:
- Spectral normalization effectively improves speech understanding for cochlear implant users when dealing with less-intelligible talkers.
- This technique can help compensate for acoustic variability across different talkers, suggesting potential for personalized CI patient calibration.
